May 2026 | Management

Your AI Investment Isn’t Paying Off Yet. Here’s Why.

Most large organizations have crossed a threshold. Since the launch of ChatGPT in late 2022, AI has moved from something companies were considering to something they were officially committed to, with large budgets and bullish predictions. The question stopped being whether AI would transform business and became when the results would arrive.

The answer? They haven’t — at least not yet. Studies from BCG, McKinsey, and MIT tell the story: fewer than 1 in 10 companies are capturing meaningful AI value at scale. The gap between ambition and impact is wide, and most organizations are still trying to figure out how to close it.

Wharton’s Executive Education program Business Model Innovation in the Age of AI was built for this moment, though its foundations predate the AI era. Now in its seventh year, the program gives senior leaders the frameworks to rethink how their organizations create, deliver, and capture value, drawing on principles that have endured across successive waves of disruption, from the platform era through COVID and into the age of AI. Its academic director, Wharton Professor Serguei Netessine, has been making the same core argument since the program launched in 2018: the obstacle to innovation is never really the technology. It’s the organization.

“AI alone doesn’t transform industries,” Netessine says. “Transformations happen when companies combine technology with innovative business models. That’s what Uber did. It didn’t invent GPS or mobile payments. It combined existing technologies in a new model and changed everything. The question for every organization right now is how to do the same with AI.”

The Messy Middle

Jeremy Korst knows that question from the inside. A Wharton MBA alumnus, member of Wharton Executive Education’s Board of Advisors, and co-author of the Wharton School and GBK Collective’s annual Enterprise AI Adoption Study, Korst led a session in the program on the human and organizational dimensions of AI transformation. He came with fresh data and a diagnosis that resonated immediately with participants.

The study surveys business leaders at U.S. companies with revenues above $50 million. The headline numbers look encouraging: more than 80 percent of leaders report weekly AI use, and 74 percent report perceived positive returns on early deployments. But when Korst and his co-authors looked beneath the averages, they found a fault line running through the leadership ranks of most organizations.

Senior executives and middle managers are not operating in the same reality. 45 percent of executives report significant ROI from their initial AI investments. For middle managers, that number drops to 27 percent. On the question of whether their organization is adopting AI faster than competitors, 56 percent of executives say yes. But only 28 percent of middle managers agree. Korst calls the space between those two realities the “messy middle,” and his session was built around helping participants understand why it exists and what to do about it. “The challenge for most organizations right now is getting from AI aspiration to AI actual,” Korst says. “Most of them haven’t made that leap yet.”

There is a compounding challenge underneath this. Middle managers are already overworked, and AI transformation requires them to rethink workflows, reskill team members, and run test-and-learn pilots, all before the promised efficiencies materialize. “Even if we have the best idea and the best technology,” Korst told participants, “if we can’t get the right people to adopt and act on it, the rest is irrelevant.”

Bringing AI to Work vs. Putting AI to Work

One of Korst’s most useful contributions to the session was a distinction that participants found immediately applicable: the difference between bringing AI to work and putting AI to work for you.

Bringing AI to work means adopting it within existing workflows: using it to manage email, summarize documents, or assist with routine tasks. That, Korst says, is the first checkbox. “But to truly capture the value, you have to put AI to work for you and your organization. And that takes rethinking how you work, how you develop products, how you bring them to market, and how you support your customers. That’s a transformation, and the technology isn’t going to do it for you.”

That distinction connects directly to the program’s central framework. Netessine’s approach to business model innovation pushes participants to inventory their entire value chain — from upstream product development and operations to logistics, delivery, and customer support — and identify where AI can create genuine differentiation. The key insight is that there is no universal AI playbook. How an organization should deploy AI depends on its strategy, and that strategy should look different from a competitor’s. “The essence of business model innovation is differentiation,” says Netessine. “The way you adopt AI should reflect that.”

To help participants move from strategy to application, Netessine offers what he calls the Four A’s framework, a risk-based approach to deploying AI across a spectrum from assistant, where AI helps complete a task, to autonomous, where AI takes over the task entirely. The framework gives leaders a practical way to think about where AI belongs in their operations, calibrated to the risk, frequency, and complexity of each decision.

Not All New

One of the more unexpected moments in the program came when participants began recognizing patterns from their own professional histories in what they were hearing about AI.
Korst, whose career includes senior executive roles at Microsoft and T-Mobile, had just heard Wharton Professor Rahul Kapoor walk through lessons from the mobile and PC industries, including how platform ecosystems formed, how value chains shifted, and how incumbents responded to disruption. “He was spot on,” Korst says. “Those are exactly the lenses we should be using to look at what’s happening now.”

Netessine reinforces this perspective throughout the program. The same principles that governed innovation in previous technological eras — the importance of ecosystems, the dynamics of platform competition, the challenge of organizational change — apply to AI. That framing, both instructors found, gave participants something valuable: not a reason to be complacent, but a foundation for action.

“By the end of the day, participants were recounting their own experiences and saying, ‘This is similar to back when,’” Korst recalls. “They were already putting that lens on, asking how they could learn from the past and apply it to now.”

Netessine is clear that this doesn’t mean AI is simply more of the same. “There are things about AI that are genuinely new,” he says. “The speed of change, the potential for autonomous decision making, the scale of impact on white-collar work, these are new. The question is which frameworks from the past still apply, and where we need entirely new thinking.”

From Aspiration to Action

For participants, the program’s value was less about understanding AI in the abstract and more about leaving with something they could use. Netessine has seen that pattern repeat across seven years of running the program. Participants come in focused on the technology and leave focused on the organization. They come in asking what AI can do and leave asking what their business model needs to become. The questions haven’t changed much since 2018; what has changed is the urgency.

“The core challenge has always been the same,” Netessine says. “Large organizations struggle to innovate. Cultural barriers, slow adoption, lack of processes for reviewing and changing business models — these are not new problems. What AI has done is raise the stakes for solving them.”

The program is still doing what it was designed to do: giving leaders the frameworks to close the gap between where their organizations are and where they need to be. The technology will keep changing. The organizational challenge is the constant.

---

Share This

Subscribe to the Wharton@Work RSS Feed